Using double-blind enriched discontinuation designs, early lithium maintenance studies demonstrated that 60-80 percent of patients with bipolar disorder had a satisfactory clinical response. However, these older studies excluded lithium-refractory subgroups such as all bipolar II disorder. When the response rate of lithium is considered across the wide spectrum of bipolar disorders, it may approach 50 percent. A large subgroup of lithium-resistant patients is rapid-cyclers. Fourteen to 53 percent of patients with bipolar disorder cycle rapidly and the majority is bipolar II and female. Seventy-two to 82 percent of rapid-cyclers exhibit poor response to lithium. Thus, a substantial percentage of poor response to lithium is attributable to rapid cycling. Preliminary data from MH-50165 suggest that combination therapy with lithium and divalproex results in marked antimanic, but modest antidepressant efficacy in patients with rapid cycling; 75 percent of non-response is attributable to resistant depression. More broadly effective regimens are needed. To address this need, we propose to compare the safety and efficacy of the triple regimen lithium, divalproex, and lamotrigine to the double-regimen lithium and divalproex. Cancer chemotherapy trials have employed random assignment to parallel arms to compare the safety and efficacy of 2 partially effective, chemotherapeutic agents to 3 partially effective agents; the evidence from MH-50165 suggests that rapid-cycling is sufficiently treatment-refractory to merit a study employing similar methodology. The primary objective of this pilot study is to conduct an exploration of the relative efficacy of the triple vs. double regimen as "first-line" therapy among unselected patients in the acute and continuation outpatient management of depression, hypomania, or mania. The proposed study will be used as pilot data for a future full-scale, STEP Prograin multicenter maintenance trial. In this revised competitive renewal, we propose a single-center, STEP-Program affiliated, 6-month, randomized, double-blind, balanced parallel-group comparison of these two regimens of combination therapy in 90 patients with rapid-cycling bipolar disorder presenting depressed over a three-year period.